Prognostic Factors, Surgical Outcomes, and Guidelines for Managing Metastatic Spine Cancer

Published on 02/04/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2772 times

Chapter 108 Prognostic Factors, Surgical Outcomes, and Guidelines for Managing Metastatic Spine Cancer

Alexander Hadjipavlou, Kyriakos Kakavelakis, Panagiotis Korovessis, James W. Simmons, II

Vertebral metastases constitute a major event in the course of many tumors. Quite often revealing the metastatic nature of the disease, they also constitute a new crossroads in treatment orientation and provoke a marked emotional modification in the caregiver-patient relationship.

The spine is the most common site for skeletal metastasis.¹ Approximately 5% to 10% of all cancer patients develop spine metastasis.^2,³ Although Jaffee¹ reported in 1958 a 70% incidence of vertebral metastasis, others asserted that spinal involvement ranges from 33.4% to 56%.⁴^–⁶ Spinal cord compression occurs in 10% to 25% of treated patients with preexisting spinal metastasis.^2,⁶^–¹² It is well established that if metastatic vertebral disease is left untreated, it may lead to paraplegia,¹³^–¹⁶ which is the most dreadful complication of metastatic vertebral disease. In this situation, surgery plays an important role. Several surgical options exist; however, their indications and outcomes must be scrutinized for better medical care. The discovery of vertebral metastases intensifies an already complex situation requiring highly technical management, sustained coherence of the entire care-provider team, and their awareness of the patient’s increased need to be heard, informed, and supported.

Constructing reasonable guidelines for monitoring patients with vertebral metastases, it is important to understand the patient’s expectations after surgery and what constitutes prognostic factors.

Surgical Outcomes

The following are essential when evaluating surgical outcomes for vertebral metastases:

• Overall neurologic status

• Complications (including immediate postoperative mortality)

• Reoperation rate

• Overall patient satisfaction

According to published reports, the anticipated improvement after surgery is expected to range from 55% to 87% of cases as a result of overall upgrade of neurologic deficit (Table 108-1); 36% to 89% because of restoration of sphincter control (Table 108-2); 47% to 100% due to pain relief (Table 108-3); and 40% to 100% as a result of restoration of ambulation (Table 108-4). Apparently restoration of ambulation after dorsal decompression alone fares worse. The anticipated survival rates are depicted in Table 108-5. The complication rates range from 12% to 32% (Table 108-6). In our series of 70 patients, the complication rate was 25%, including two deaths. In one case, surgery was aborted, while in progress, because of massive intraoperative bleeding, despite preoperative transarterial embolization of the region.¹⁷ Other complications were profuse bleeding in nine patients, infections in three, wound dehiscence in three, and incomplete surgery in three patients. The 4- to 6-week mortality rates after surgery, reoperation rates for complications, and local recurrences after surgery are listed in Tables 108-7, 108-8, and 108-9, respectively. Local tumor recurrence is more likely to occur in patients with certain tumors and particularly in those for whom chemotherapy and radiation therapy are ineffective. This is particularly the case in such cancers as renal cell carcinoma, in which the reported recurrence is expected to be as high as 40% to 50%.¹⁸^–²¹ The overall patient satisfaction was rated in the range of 55% to 80%.^18,²²

TABLE 108-1 Reported Overall Neurologic Improvement after Surgery

Study	Improvement
Kostuik J et al.⁶⁸	55%
Ernstberger T et al.⁸⁰	57%
Weigel B et al.¹⁸	58%
King GJ et al.¹⁹	60%
Villavicencio AT et al.⁸¹	60%
Rompe JD et al.⁸²	63%
Atanasiu JP et al.³⁰	64%
O’Neil J et al.³²	68%
Harrington KD³⁷	68%
Hatrick N et al.⁸³	69%
Tomita T et al.⁷²	73%
Hammerberg KW⁶⁶	74%
Bauer HCF⁸⁴	76%
Gokalsan ZL et al.⁴⁶	76%
Onimus M et al.⁸⁵	79%
Hosono N et al.⁸⁶	81%
King et al.¹⁹	88%
Hertlein H et al.⁸⁷	88%
Shimizu K et al.³⁴	87%

TABLE 108-2 Restoration of Sphincter Control after Surgery

Study	Successful Restoration
King GJ et al.¹⁹	36%
Kocialkowski A, Webb JK⁸⁸	43%
Kostuik J et al.⁶⁸	47%
Sinardet D et al.⁸⁹	51%
Tomita T et al.⁷²	77%
Klekamp J, Samii H³¹	89%

TABLE 108-3 Pain Relief

Study	Success Rate
Sinardet D et al.⁸⁹	47%
Touboul E et al.⁹⁰	71%
Hirabayashi H et al.²⁵	77%
Kocialkowski A, Webb JK⁸⁸	79%
Kostuik J et al.⁶⁸	81%
Sundaresan N et al.⁹¹	85%
Ernstberger T et al.⁸⁰	85%
Villavicencio AT et al.⁸¹	85%
King GJ et al.¹⁹	88%
Weigel B et al.¹⁸	89%
Hatrick N et al.⁸³	90%
Hammerberg KW⁶⁶	91%
Gokalsan ZL et al.⁴⁶	92%
Hertlein H et al.⁸⁷	92%
Harrington KD³⁷	94%
O’Neil J et al.³²	94%
Hosono N et al.⁸⁶	94%
Atanasiu JP et al.³⁰	95%
Tomita T et al.⁷²	95%
Shimizu K et al.³⁴	100%
Bilsky MH et al.⁹²	100%

TABLE 108-4 Restoration of Ambulation after Ventral or Dorsal Approach

Study	Successful Restoration
Livingston KE, Perrin RG⁹³	40%
Kostuik J et al.⁶⁸	44%
Gokalsan ZL et al.⁴⁶	51%
Sinardet D et al.⁸⁹	52%
Tomita T et al.⁷²	60%
Patchel RA et al.⁹⁴	62%
Hosono N et al.⁸⁶	64%
King GJ et al.¹⁹	64%
Weigel B et al.¹⁸	70%
Onimus M et al.⁸⁵	70%
Hammerberg KW⁶⁶	76%
Hatrick NC et al.⁸³	78%
Siegal T, Siegal T⁹⁵	80%
O’Neil J et al.³²	82%
Villavicencio AT et al.⁸¹	100%
Sundaresan N et al.^91,⁹⁶	100%
Restoration of Ambulation after Laminectomy Alone
Klekamp J, Samii H³¹	23%
Kennady JC, Stern WE⁹⁷	23%
Schoeggl A et al.⁹⁸	33%
Wright RL⁹⁹	35%
White WA et al.¹⁰⁰	36%
Siegal T, Siegal T⁹⁵	47%

TABLE 108-5 Anticipated Survival Rates after Surgery

Study	Survival Rate
3 month
Kocialkowski A, Webb JK⁸⁸	50.0%
Wai EK et al.⁵²	76.0%
6 month
Klekamp J, Samii H³¹	58.8%
Rompe JD et al.⁸²	72.0%
Hosono N et al.⁸⁶	78.0%
9 month
Falicov A et al.³³	50.0%
12 month
Kocialkowski A, Webb JK⁸⁸	25.0%
Klekamp J, Samii H³¹	48.8%
Rompe JD et al.⁸²	50.0%
Villavicencio AT et al.⁸¹	65.0%
Gokalsan ZL et al.⁴⁶	62.0%
19 month
Rompe JD et al.⁸²	4.7%
60 month
Klekamp J, Samii H³¹	5%
Median month
Schoeggl A et al.⁹⁸	6.5%
Sioutos PJ et al.³⁵	10%
Hirabayashi H et al.²⁵	10.6%
Lewandrowski KU et al.¹⁰¹	14%
Sundaresan N et al.¹⁰²	30%
Mean month
Chataigner H, Onimus M²⁰	8%
Hertlein H et al.⁸⁷	8.5%
O’Neil J et al.³²	9%
Livingston KE, Perrin RG⁹³	9%
Atanasiu JP et al.³⁰	11%
Ernstberger T et al.⁸⁰	15.6%
Weigel B et al.¹⁸	13.1%
Wise JJ et al.⁶⁷	15.9%
Cahill DW, Kumar R¹⁰³	15.9%
Kocialkowski A, Webb JK⁸⁸	12.0%

TABLE 108-6 Complication Rates

Study	Percentage Experiencing Complications
Bilsky MH et al.⁷	12%
Wai EK et al.⁵²	12%
Hosono N et al.⁸⁶	16%
Weigel B et al.¹⁸	19%
Jansson KA, Bauer HC¹⁰⁴	20%
Wise JJ et al.⁶⁷	25%
Klimo P Jr et al. (review of 24 surgical reports)¹⁰⁵	23%
Falicov A et al.³³	29%
Gokalsan ZL et al.⁴⁶	32%

TABLE 108-7 Postsurgery Mortality Rates (4 to 6 weeks)

Study	Mortality Rate
Gokalsan ZL et al.⁴⁶	3.0%
Klimo PJr et al.¹⁰⁵	6.3%
Wise JJ et al.⁶⁷	6.0%
Siegal TJ et al.¹⁰⁶	6.0%
Weigel B et al.¹⁸	7.0%
Sundaresan N et al.⁹⁶	8.0%
Harrington KD³⁷	8.0%
Sinardet D et al.⁸⁹	9.0%
Sioutos PJ et al.³⁵	11.0%
Turner PL et al.⁷⁸	10.0%
Hertlein H et al.⁸⁷	10.0%
Chataigner H, Onimus M²⁰	10.2%
Bilsky MH et al.⁹²	12.0%
Jansson KA, Bauer HC¹⁰⁴	13.0%
Kocialkowski A, Webb JK⁸⁸	16.0%
Fidler MW¹⁰⁷	18.0%

TABLE 108-8 Reoperation Rates for Complications or Local Recurrences

Study	Percentage Undergoing Reoperation
Weigel B et al.¹⁸	7.0%
Jansson KA, Bauer HC¹⁰⁴	10.2%
Hertlein H et al.⁸⁷	12.1%
Bauer HCF⁸⁴	24.0%
Chataigner H, Onimus M²⁰	15.8%
King GJ et al.¹⁹	56.2%

TABLE 108-9 Local Recurrences after Surgery

Study	Percentage Experiencing Recurrence
Falicov A et al.³³	3.5%
Review of 9 surgical reports	8.0%
Klimo P Jr et al.¹⁰⁵
Bilsky MH et al.⁹²
Foumey DR et al.¹⁰⁸
Hammerberg KW⁶⁶
Hosono N et al.⁸⁶
King GJ et al.¹⁹
Muhlbauer M et al.¹⁰⁹
Siegal T et al.¹⁰⁶
Weigel B et al.¹⁸
Chataigner H, Onimus M²⁰	8.4%
Nazarian S¹¹⁰	11.0%
King GJ et al.¹⁹	12.1%
Hirabayashi H et al.²⁵	21.0%
Weigel B et al.¹⁸	22.0%
Bridwell KH et al.⁷¹ (after laminectomy)	31.6%
Sundaresan N et al.¹⁰²	32.0%
Hertlein H et al.⁸⁷	49.0%

Therefore, it is reasonable to assume that these data can be used as the basis for what should be considered acceptable outcomes. The reported wide variation of survival rates after surgery probably reflects nonuniform representation of aggressive tumors in different series.

Prognostic Factors

It has been observed that certain factors, irrespective of treatment selection, may influence the prognosis when malignant tumors develop spinal metastases. The original claim that the histopathology of the tumor had no bearing on the ultimate prognosis ²³ has not been duplicated by others.

Origin of Tumor

The origin of the primary tumors was postulated by several authors to influence the prognosis. Thus, metastatic vertebral tumors from unsuspected adenocarcinoma, stomach, esophagus, pancreas, and lungs, portend the worst possible prognosis, whereas carcinoid, thyroid, breast, prostate, and myeloma carry the best prognosis. Distinction should also be made between the different types of lung cancer.²⁴ In general, non–small cell lung cancer (NSCLC) patients with spinal metastases had survival rates of 37.1%, 14.6%, and 2.1% at follow-up of 6 months, 1 year, and 2 years, respectively. The median survival time was 4.5 months, and the mean, 6.2 months. However, for small-cell lung cancer (SCLC), the corresponding survival rates were 36.8%, 5.3%, and 0% at 6 months, 1 year, and 2 years follow-up, respectively. In both NSCLC and SCLC with spinal metastases, the presence of hypercalcemia or hypoalbuminemia was indicative of a gloomy prognosis, with a survival period of less than 3 months.²⁴

The aggressiveness of cancer can also be classified as slow growth (breast, thyroid, prostate), moderate growth (kidney, uterus), and rapid growth (lung, stomach, liver, colon, unknown).²⁵^–²⁹

Concurrent Visceral Metastases

The influence on survival with visceral metastasis was also observed by Tokuhashi et al.²⁸ and Tomita et al.²⁹ The average survival with no visceral metastases was 36.8 months (range, 5–84 months), treatable visceral metastases, 16.5 months (4–31 months), and untreatable visceral metastases, 8.9 months (range, 1–24 months). The grave prognostic factor of extraskeletal metastatic lesions on the survival rate is also in agreement with Weigel et al.,¹⁸ who demonstrated that a patient survival period without extraskeletal metastases at surgery was significantly longer than with extraskeletal metastases (23.5 months vs. 5.8 months; P < .0001).

Location of Spinal Lesion

The location of the metastatic lesion was also implicated as a survival factor. Atanasiu et al.³⁰ reported that lesions in the upper cervical spine had an adverse effect on life expectancy, with average survival of 1.8 months; however, this was challenged by Klekamp and Samii,³¹ who found no differences in prognosis between the upper and lower spine.

Number of Affected Vertebrae

The number of affected vertebrae also had an influence on survival. Patients with single-level involvement had a better survival rate (average survival, 12.9 months) than did patients with multiple-level involvement (average, 7.9 months),^28,^32,³³ even after effective surgical treatment.³⁴

Neurologic Deficit

Neurologic dysfunction was observed to be associated with survival rates. Some believe that ambulatory patients survive longer than nonambulatory patients with sphincter incontinence.^19,³⁵ The state of sphincter incontinence as a bad prognostic sign is also emphasized by some.³⁶ The rate of neurologic deterioration was related to prognosis; slow progression had a good prognosis as opposed to rapid progression, which was correlated with dismal results. The latter was compounded when the onset of treatment was delayed.^36,³⁷ However, the notion of preoperative intact neurologic status (ambulatory vs. nonambulatory) as a predictor of longer survival was challenged by others.^18,³⁸

Likewise, timing for adjunctive radiation treatment (preoperative radiation vs. postoperative radiation) was also questioned as a significant prognostic predictor for outcomes.³⁹ Weigel¹⁸ and Hirabayashi et al.²⁵ emphasized that postoperative ambulation definitely has a positive prognostic influence on survival, rather than preoperative ambulation.

Timing of Surgery

The timing of treatment can also influence the results. Some authors ^36,³⁷ observed that delayed onset of treatment may compromise the effectiveness of neurologic recovery. More specifically, Fürstenberg et al.⁴⁰ addressed the timing of surgery and concluded that the effectiveness of decompression, when undertaken less than 48 hours after the development of symptoms, was significantly better for neurologic recovery (71.4%, unchanged 28.6%), compared with those with delayed surgical treatment (improvement 28.6%, unchanged 42.8%, deterioration 28.6%). Furthermore, they noted that normal bladder function may be considered a good prognostic factor for neurologic recovery after appropriate decompression.

Age

Weigel et al.¹⁸ also reported that the patient age can be considered as a prognostic factor for survival, because patients younger than 60 years survived significantly longer than older patients (20.1 months vs. 6.2 months; P = .028). Age as a prognostic factor has not been incorporated into the Tokuhashi scoring system and needs to be duplicated by others.

More Than Two Prognostic Factors

A report indicates that when more than two of the “prognostic factors” are present, they have a compounding adverse effect on survival.³⁵ For example, lung cancer, neurologic deficit, and involvement of multiple vertebrae in the same patient would have a greater adverse effect on survival than the combined effect of the same type of lung cancer involving a single vertebra and without neurologic deficit. However, these findings have not been substantiated by others. When estimating survival, one should take into consideration that, invariably, most of the data from different reports were based on metastatic tumor of the spine that became symptomatic.

Clinical practice in surgical results and prognostic factors of metastatic spinal tumors has not undergone adequate scrutiny. The surgical information provided in this article must be available and known to everyone concerned with spinal tumors (e.g., physicians, patients, insurance) to form the basis of what should be considered acceptable surgical outcomes. Inferior results, therefore, are not acceptable, and well-designed guidelines should strive to achieve the best possible outcomes.

Instruments for Outcomes Assessment

As surgical procedures to treat vertebral metastases have increased, a greater interest has been expressed in the overall quality-of-life assessment and outcomes studies. However, few published reports exist on the assessment of these patients’ health-performance status after surgery.

Patients with metastatic spinal disease often suffer from a multitude of concurrent clinical problems. Quality of life may be compromised by the effects of chemotherapy or surgery in conjunction with radiation therapy, as well as by the effects of the disease itself. Furthermore, survival may also be limited, and thus long-term follow-up is not feasible for this patient population.

Patient-based outcome studies have been recently introduced to assess the results of treatment and to provide patients, physicians, insurance providers, and government agencies with a tool with which to evaluate the different treatment methods. The objective is to measure how well the results approach the goals of treatment (i.e., restore function and control pain).

Evaluation of patient functional status should be based on (1) pain assessment, (2) profile of mood states, and (3) overall performance status. Because performance depends not only on pain and disability but also on neurologic status, a detailed neurologic assessment is an integral part of the evaluation of these patients and provides reliable and straightforward clinical documentation. However, the assessment of pain disability and bodily functional performance of the patient can be appreciated only through psychometric instruments, which should be simple and reproducible.⁴¹

Because treatment is seldom curative in metastatic disease,⁴² the treating physician should keep in mind that patients with vertebral metastasis differ from noncancer patients. Therefore, the sensitivity of these instruments is limited to monitoring changes of the patient’s feelings of well-being.^43,⁴⁴

Pain is an alarming consequence of vertebral metastatic disease. Specific efforts should be made to avoid crude assessment and try to quantify and locate the pain. The McGill Pain Questionnaire is a sensitive tool. A visual analogue scale (VAS) for pain measurement combined with a pain diagram that indicates the site of pain and its spread is a simple and reliable method for quantifying and depicting pain. The VAS, although a subjective assessment of pain, has been shown to be more reliable than other types of pain assessment.⁴⁵

A distinction should be made between radicular pain and axial pain. Even the surgical approach may be responsible for severe pain. A report ⁴⁶ detailing outcomes of pain demonstrated that in 72 patients who underwent a ventral approach through thoracotomy, 90% had postoperative pain (only 23% achieved complete resolution, 60% had significant improvement, and 8%, no change or worsening of thoracotomy pain). Quantification of pain is useful in monitoring the effectiveness of pain management, as well as in helping establish pain-control goals for the individual patient.⁴⁷

Cancer patients often have emotional and psychological problems. Because anxiety and depression have a direct correlation with pain intensity, these two factors may need to be addressed. The Zung Self-Rating Anxiety Scale ⁴⁸ and the Hamilton Depression Scale⁴⁹ are suitable instruments in this situation. The Memorial Pain Assessment Card (MPAC)⁵⁰ (Fig. 108-1) is a modified visual analogue with multidimensional characteristics, practically equivalent to a full assessment combining the McGill Pain Questionnaire, the Hamilton Depression Rating Scale (HDRS), and the Zung Self-Rating Anxiety Scale.

FIGURE 108-1 The Memorial Pain Assessment Card.

(From Fishman B. Pasternak S, Wallenstein SL, et al: The Memorial Pain Assessment Card: a valid instrument for the evaluation of cancer pain. Cancer 60:1151, 1987.)

The Edmonton Symptom Assessment Scale (ESAS) is a validated patient-centered questionnaire designed specifically to address the overall quality of life in patients with terminal cancer.^43,^44,⁵¹ ESAS measures nine domains (pain, tiredness, nausea, depression, anxiety, drowsiness, appetite, well-being, shortness of breath, and other problems). After surgery, the largest improvement is reported in the pain domain.⁵²

The 36-Item Short Form Health Survey (SF-36) by Townsend et al. is a physician-determined assessment of the patient’s physical disability. Patients are classified as having (1) normal, pain-free function, (2) normal function with pain, (3) significantly limited function requiring some type of assistance (e.g., a walker, cane), and (4) nonfunctional (e.g., wheelchair bound, bedridden).⁵³

Other validated outcome measures in patients suffering from vertebral metastasis are the Health Utility Index Mark 3 (HUI-MARK 3; Table 108-10),⁵⁴ European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30),⁵⁵ Eastern Cooperative Oncology Group (ECOG; Table 108-11),⁵⁶ EuroQol 5d (EQ-5D; Fig. 108-2),⁵⁷ and Karnofsky Performance Status (PS; Table 108-12), which have been successfully applied in patients with vertebral metastases.⁵⁸

TABLE 108-10 Human Utility Index Mark 3

Scale	Description
1	FREE OF PAIN AND DISCOMFORT
2	MILD TO MODERATE PAIN Pain does not prevent activity
3	MODERATE PAIN Pain prevents few activities
4	MODERATE TO SEVERE PAIN Pain prevents some activities
5	SEVERE PAIN Pain prevents most activities

TABLE 108-11 ECOG Performance Status

Grade	ECOG
0	Fully active, able to carry on all predisease performance without restriction
1	Restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, e.g., light housework, office work
2	Ambulatory and capable of all self-care but unable to carry out any work activities; up and about more than 50% of waking hours
3	Capable of only limited self-care; confined to bed or chair more than 50% of waking hours
4	Completely disabled. Cannot carry on any self-care. Totally confined to bed or chair.
5	Dead

ECOG, Eastern Cooperative Oncology Group.

FIGURE 108-2 EQ-5D self-classifier (that consists of five dimensions) and visual analogue scale to enable the respondent to provide a health self-rating.

TABLE 108-12 Rating of Karnofsky Performance Status

Rating	Status	Description
0	Dead	Unable to care for self; requires equivalent of institutional or hospital care; disease may be progressing rapidly.
10%	Moribund Fatal processes progressing rapidly
20%	Very sick; hospitalization necessary Active support treatment necessary
30%	Severely disabled; hospitalization is indicated, although death not imminent
40%	Disabled Requires special medical care and assistance	Unable to work; able to live at home and care for most personal needs; varying amount of assistance needed.
50%	Requires considerable assistance and frequent medical care
60%	Requires occasional assistance, but is able to care for most needs
70%	Cares for self Unable to carry on normal activity or to do active work
80%	Normal activity with effort Some signs or symptoms of disease	Able to carry on normal activity and to work; no special care needed.
90%	Able to carry on normal activity Minor signs or symptoms of disease
100%	Normal No complaints No evidence of disease

From Karnofsky DA: Clinical evaluation of anticancer drugs: cancer chemotherapy. GANN Monogr 2:223, 1967.

The overall distribution of HUI-MARK 3 utility calculates quality-adjusted life years (QuaLY) during the 1-year postoperative period.³³

The EORTC QLQ-C30 measures quality of life and has one global health status (QL2), five functional scales, physical functions (PF2), role functions (RF2), emotional function (EF), cognitive functions (CF), and social functioning (SF), as well as nine symptom scales.

The ECOG grading system or the Karnofsky PS is used to assess performance status. The ECOG (see Table 108-11) performance status⁵⁶ measures functional performance on a scale from 0 to 4 and is a good index of assessing the preoperative functional status.³³

The Karnofsky PS details the condition of the patient’s health status ⁵⁸ (see Table 108-12). Scoring between 0 and 10% indicates poor health and total dependency on others, from 50% to 70% indicates an inability to work and requires assistance, and 80% to 100% represents the ability to carry on with normal activities. The Karnofsky PS was successfully applied by Tokuhashi et al.⁵⁹ in clinical guidance for predicting cancer patient outcomes.

Frankel et al.⁶⁰ and the American Spinal Injury Association (ASIA) impairment scores,⁶¹ although nonspecific enough to describe nuances of the physical state of locomotion, grade the functional outcomes for neurologic assessment. Four other scales also address the state of ambulation^60,⁶²^–⁶⁴ and can be used as complementary to the ASIA or Frankel scales, such as combining the Cooper and the Frankel grading systems.

Oswestry Disability Index (ODI) and SF-36 are simple tools designed to quantify disability originating from back pain. The effectiveness of these tools has not been tested in patients with metastatic spinal disease.

Surgical Guidelines

The rationale of guidelines is to achieve better treatment outcomes. Their scientific strength depends on how well evidence-based knowledge is used.⁶⁵ It is reasonable to assume that consensus-based guidelines supported by available evidence are more convincing than individual ones. Unfortunately, no such reported guidelines are available for the management of metastatic spinal tumors. Certain published guidelines are based on individual experiences without enough clinical evidence to support their effectiveness. These guidelines do not cover an integrated approach for the management of metastatic spinal tumors but are concerned with different aspects of the problem.

In 1990, Tokuhashi et al.⁵⁹ devised a guideline scoring system for the preoperative evaluation of metastatic vertebral tumor, based on the general condition of the patient, as graded by the Karnofsky PS⁵⁸ and five of the following prognostic factors: (1) number of extraspinal bone metastases, (2) number of metastases in the spine, (3) metastases to the major internal organs (lungs, liver, kidneys, and brain), (4) primary site of the cancer, and (5) severity of spinal cord impairment. Table 108-13 demonstrates the Tokuhashi prognostic scoring system for preoperative evaluation of the metastatic spinal lesions. Each parameter has a score ranging from 0 to 2 points (maximum score 12 points) that is related to the patient’s prognosis: the higher the score, the better the prognosis.

TABLE 108-13 Tokuhashi System for Preoperative Evaluation of Metastatic Spine Tumors

Parameters	Score
General condition (performance status)	0, 1, 2
Extraspinal bone metastatic foci	0, 1, 2
Number of metastases in the spine (vertebrae)	0, 1, 2
Metastasis to major internal organs	0, 1, 2
Primary site of the cancer (histopathology)	0, 1, 2
Spinal cord palsy	0, 1, 2
Indications for Type of Surgery	Score
Excisional surgery	>9
Palliative surgery	<5

Rating the preoperative health status is as follows: the score is 0 for poor condition (PS values, 10–40%); 1 for moderate (PS values, 50–70%), and 2 for good (PS values, 80–100%). For extraspinal bone metastatic foci more than three, the score is 0; between one and two metastatic foci, the score is 1; and for nonmetastatic lesions, 2. For the number of metastases in the spine, the score is 0 for more than three vertebrae; 1 for two metastases; and 2 for no more than one vertebra. For metastasis to a major internal organ, the prognosis depends on the resectability of the tumor: 0 for unremovable lesions; 1 for removable lesions, and 2 for no metastasis. For the primary site of the cancer, the score is: 0 when the expected survival is less than 3 months (lung and stomach); 1 for expected survival rate of 3 to 12 months (kidney, liver, uterus, others, unidentified), and 2 for expected survival rate of more than 12 months (thyroid, prostate, breast, rectum). Finally, the severity of spinal cord palsy was rated as 0 for Frankel grade A and B (complete paraplegia) 1 for Frankel C and D (incomplete paraplegia), and 2 for Frankel E (no paraplegia), respectively.

Excisional surgery should be performed for patients who score above 9 points, and palliative surgery for those who score less than 5 points. The patients who score between 5 and 9 points fall into a gray zone.

In 1997, Tokuhashi et al.²⁸ reported that the reliability for predicting the prognosis was 63.3%, tested in 128 patients with metastatic spinal tumors.

The original Tokuhashi scoring system was tested by Enkaoua et al.²⁶ for three types of metastatic tumors: renal cell carcinoma, thyroid cancer, and unknown primary cancer. The median survival period was 5.3 months for a score of 7 or less and 23.6 months for a score of 8 or more. For cancer of an unknown primary site, the score should be 0 because the prognosis is invariably gloomy.

Subsequently, Tokuhashi et al.²⁸ revised the original scoring system to improve its accuracy by providing more detailed prognostic values based on the origin of the primary tumor. The score was modified to range between 0 and 5 points. The score was designated as 0 for metastases originating in the lungs, osteosarcoma, stomach, bladder, esophagus, and pancreas; 1 for metastases from the liver, gallbladder, and unidentified sources; 2 for metastasis from other sites; 3 for renal or uterine malignancies; 4 for metastasis from the rectum; and 5 for tumors from the thyroid, breast, or prostate or a carcinoid tumor. Therefore, in the revised system, a total of 15 scores can be obtained. In the patients with a total score of less than 8 or multiple vertebral metastases, Tokuhashi et al. indicate conservative or palliative procedure (predicted survival period is less than 6 months), whereas they advocate excisional procedure when the score is 12 or more (predicted survival period, 1 year or more) and for scores between 9 and 11 (predicted survival period, 6 months or more). According to the authors, the consistency rate between the criteria for predicted prognosis and the actual survival period was high in patients within each score range (0–8, 9–11, 12–15), 86.4% in 118 patients evaluated prospectively after 1998, and 82.5% in all 246 patients evaluated retrospectively. No reports in the literature indicate that the actual value of each scoring point (0, 1, and 2) in each factor has been sufficiently scrutinized. Ogihara et al.²⁴ reported that the Tokuhashi preoperative scoring system for patients with lung cancer did not correlate well with the survival period. The authors found that only the performance status ECOG (PS; see Table 108-11),⁵⁶ which is similar to the Karnofsky test ,was correlated with the survival rate. However, this test was predictive only when applied postoperatively. The average survival rate for patients with a postoperative PS of 0 to 2 was 13.9 months, whereas that for patients with a postoperative PS of 3 or 4 was 2.0 months.

By using the Karnofsky PS, Weigel et al.¹⁸ found that the postoperative general state of health was 55%, a state of health that lies halfway between total dependency (50%) and partial dependency on others (60%). No one achieved 100%. The best scores were achieved with breast tumors (80% on the index); the second best results were seen in renal cell carcinoma patients (30% of the patients scoring 80–90%); and the poorest scores were seen in patients with melanoma (mean index, 48%), tumors of other sites (43%), and lung tumors (38%).

Chantaigner et al.²⁰ asserted that when deciding on a surgical option in non-neurologic patients, the Karnofsky PS alone was more effective than the Tokuhashi score, which also includes the Karnofsky PS.

Some investigators are more selective than others in applying more vigorous selection in their inclusion criteria for surgery. Therefore, the overall survival period after surgery is not accurate. In some series, surgery is not indicated when the estimated survival rate is 3 months, whereas in others, the criterion for surgery is longer than 6 months. This explains the longer survival rates with different cancer patients and particularly in lung cancer (6–12 months) in some reports.^39,⁶⁶

Harrington ³⁷ described a classification guideline system based on both bone destruction and neurologic impairment, and distinguishes five classes of vertebral metastatic lesions (Table 108-14). He suggested surgical intervention for class IV and V lesions and conservative treatment for class I and II, while for class III, which lies in the gray zone, treatment depends on various circumstances. Harrington’s³⁷ grading system is a practical guideline system for managing metastatic vertebral tumors, but it has some significant limitations: (1) it makes no recommendations for surgical options; (2) it does not take into account the biology of the tumor nor (3) the radiosensitivity of the metastatic lesion; and (4) it makes no provisions for the performance status of the patient. The applicability of the Harrington³⁷ classification system for selecting the appropriate treatment in metastatic cancer of the spine was also applied by others.⁶⁷

TABLE 108-14 Harrington Classification Based on Structural Defect and Neurologic Deficit

Class	Neurologic Status	Structural Changes
I	Not significant	No vertebral collapse
II	Not significant	Vertebral involvement without collapse or instability (lytic or blastic lesion)
III	Major (sensory or motor)	No significant bone destruction or instability (lytic tumor)
IV	Not significant	Mechanical pain from vertebral collapse ± instability
V	Major	Retropulsion of hard discovertebral elements ± kyphotic deformity

Harrington ³⁷ alluded to instability as one of the deciding factors for surgical stabilization because vertebral instability, caused by bone destruction, would provoke pain. However, he has not determined what constitutes instability in destructive metastatic lesions.

Kostuik et al.^68,⁶⁹ devised a vertebral body stability model by dividing the vertebral body into six segments. More specifically, the vertebral body is divided into an anterior column consisting of the vertebral body and a posterior column consisting of the pedicles, laminae, and spinous processes. The anterior column is further subdivided into four segments: anterior right, anterior left, middle right, and middle left. The posterior column is also subdivided into right and left segments. The spine is considered stable if no more than one or two of the six segments is destroyed. The spine is considered probably unstable if three or four of the segments are destroyed, and if five or six segments are affected, the spine is considered markedly unstable. Although this has not been biomechanically tested or clinically substantiated, it may have potential applications when describing instability in the Harrington classification.

According to Taneichi et al.,⁷⁰ impending collapse is indicated (1) when more than 50% to 60% of the vertebral body is destroyed, or (2) when 25% to 30% of the vertebral structure is destroyed when there is an associated destruction of the costovertebral junction in the thoracic spine, or (3) when the destruction in the thoracolumbar spine ventrally is 35% to 40% or 20% to 25% involves the dorsal elements. Bridwell et al.⁷¹ believe that if less than 50% of the vertebral body has been resected, vertebral body reconstruction may not be necessary.

Tomita et al.,⁷² based on Enneking’s grading of skeletal resection,⁷³ Denis’s three-column classification of fracture of the spine,⁷⁴ and Weinstein’s zone classification⁷⁵ (intraosseous, extraosseous, and distant spread), proposed a classification consisting of seven types of lesions (Fig. 108-3). This seems a reasonable classification for allowing the surgeon to choose the extent of surgical intervention—that is, radical excision versus stabilization alone.

FIGURE 108-3 Tomita classification based on vertebral spread of metastasis.

(From Tomita K, Kawahara N, Baba H, et al: Total en bloc spondylectomy for solitary spinal metastases. Int Orthop 18:291, 1994.)

The most recent topographic classification system reported by Boriani et al.⁷⁶ provides an excellent guideline for surgical excision of primary tumors of the spine and therefore is not specific for metastatic tumors.

A system that would encompass all these criteria would form a more comprehensive approach for the management of metastatic spinal malignancies (Fig. 108-4). For Harrington class I and II without significant neurologic involvement and without vertebral collapse, the management is nonsurgical and consists of chemotherapy, hormonal manipulation, and/or radiation therapy³⁷ (Fig. 108-5). However, the treatment is contingent on the origin of the metastatic tumor. For example, in renal cell carcinoma, which is resistant to radiation therapy and chemotherapy, radical excision is indicated, even in class I. For Harrington class II tumors, as in the case of multiple myeloma or small lytic metastatic lesion, considerable and immediate relief of pain can be obtained by combining chemotherapy and a cement augmentation procedure (vertebroplasty or balloon kyphoplasty). See Figure 108-6.

FIGURE 108-4 Treatment guidelines.

FIGURE 108-5 Metastatic lytic breast cancer as seen on midsagittal MRI (A) and axial CT (B). C, Assessment reveals a Harrington class II lesion with a Tokuhashi score of 12, favoring a good prognosis and nonsurgical treatment. D, After radiation therapy, the lesion healed and the patient was well and alive at 5 years follow-up.

FIGURE 108-6 A, Lytic metastatic lesion from cancer of the breast resulting in crippling back pain as seen on axial CT. B, Assessment revealed a Harrington class II lesion with a Tokuhashi score of 9. C, The patient was treated with balloon kyphoplasty and chemotherapy. At 5-year follow-up the patient was alive and well.

For Harrington category III lesions, with a neurologic deficit in the absence of major destruction of bone or spinal instability, surgery should not be entertained unless the tumor is radioresistant or unresponsive to chemotherapy.³⁷

For class IV and V, surgery is the treatment of choice. For Harrington class IV, consisting of mechanical pain caused by vertebral collapse and/or instability, with a Tokuhashi score of less than 5 with the old system, or 8 with the revised system, the indications for treatment would be either conservative treatment or palliative surgery, consisting of stabilization and/or cement augmentation (Fig. 108-7).

FIGURE 108-7 A, Lateral roentgenogram of the lumbar spine of a 67-year-old man with lung metastasis to L3 vertebra. B, Axial CT shows osteolytic lesion of the vertebral body. C, Tokuhashi score was 4, indicating a palliative surgical approach as appropriate treatment. Following minimal surgery consisting of balloon kyphoplasty and percutaneous transpedicular stabilization (D), the patient obtained immediate relief of pain and survived for 5 months, without neurologic complications or disability.

For a Harrington class IV or V with a Tokuhashi score greater than 9 (old system) or 12 (revised system) and a Tomita B score of 6 (widespread metastases), local radical excision, if feasible combined with stabilization, would be indicated (Fig. 108-8).

FIGURE 108-8 A, Lateral roentgenogram of the thoracic spine of a 72-year-old man with metastasis to the T7 vertebra from cancer of the prostate and incomplete paraplegia (ASIA C). B, Axial CT revealed osteolytic lesion in the vertebral body. C, Tokuhashi score was 9. The patient was treated with corpectomy, reconstruction with mesh cage, and posterior stabilization (D). This resulted in improvement of pain and neurologic rate in 15 months.

When the score lies in the gray zone of 5 and 9, the use of conservative or surgical treatments depends on several factors, such as the pathology of the lesion and the status of neurologic deficit, among others—factors that may influence the final decision.

For class V, characterized by intrusion of hard discovertebral elements into the spinal canal that compromise the spinal cord, the only choice is surgery. Spinal kyphotic deformity with cord compromise should be categorized as class V. In this situation, radiation has no logical place, and surgery is the treatment of choice (Fig. 108-9).

FIGURE 108-9 A 50-year-old woman with acute onset of leg weakness from metastatic breast cancer. Radiotherapy failed (neurologic deterioration). A, Sagittal CT; B, axial CT. C, Tokuhashi score was 9, indicating a good prognosis. Treatment consisted of radical corpectomy and reconstruction with titanium cage and Kaneda stabilization as seen on axial CT (D) and lateral radiograph (E). Ambulation was restored and the survival rate was over 12 months.

The ultimate intent of surgical management of metastatic spinal disease is to improve the patient’s overall quality of life, which is definitely apparent from the cited reports. The greatest effect occurs in largely the reduction of the level of the pain. In general, a minimum life expectancy of 3 to 6 months has been accepted as a prerequisite for surgery.⁶⁷ However, even with a life expectancy shorter than 3 months, surgical intervention may be justified.⁷⁷ In this situation, palliative surgery is indicated, and the anticipated surgical intervention should be performed, with the expectation of improving the remaining global quality of life, notwithstanding the morbidity of the procedure⁷⁸ (Figs. 108-10 and 108-11). Some of the procedures for alleviating pain, such as cement augmentation (vertebroplasty, balloon kyphoplasty) and neuroaugmentation (morphine pump), are associated with minimal morbidity. Surgery is also indicated for patients who have reached spinal cord tolerance after external radiation beam treatment.

FIGURE 108-10 A 60-year-old man with severe neck pain and leg weakness from metastatic lung cancer. Sagittal (A) and axial (B) MRIs. C, Tokuhashi score was 5. The patient underwent palliative treatment consisting of decompression, stabilization, and radiation (D and E). Ambulation was restored and the patient survived 11 months.

FIGURE 108-11 A 50-year-old man with severe back pain and increasing inability to walk from multiple myeloma. Lytic lesion as seen on sagittal (A) and axial CT (B). C, Assessment revealed a Tokuhashi score of 8. After a Galveston stabilization procedure (D and E), decompression, and radiation therapy, the patient’s ambulation was restored completely and he survived for more than 12 months.

It is not the purpose of this chapter to describe the available surgical techniques. In brief, the decision to use a particular approach and surgical reconstruction (ventral or dorsal) depends on the location of the tumor (e.g., vertebral body, extension into the epidural space, paraspinal region, dorsal vertebral elements) and the extent and aggressiveness of the tumor.

Summary

In this updated and upgraded version of a previously published chapter on this subject,⁷⁹ the conclusion remains that survival outcomes for managing metastatic spinal tumors are still gloomy. When assessing surgical outcomes, the following published parameters should form the baseline for acceptable outcomes: neurologic improvement (55–87%) restoration of ambulation (40–100%), restoration of sphincter control (36–89%), pain relief (47–100%); local recurrences (8–31%), complication (16–32%), reoperation rate for complications (7–24%), and patient satisfaction (55–80%). Furthermore, the reported mortality rate 4 to 6 weeks postoperatively ranges from 3% to 10.2%, whereas the mean survival is 8 to 15.9 months. Pain shows the greatest response. In addition, the prognostic factors that can influence the final outcomes are origin of tumor, visceral metastasis, number of affected vertebrae, state of neurologic dysfunction, and onset of surgical treatment. It is important to quantify the patient’s pain, psychological distress, and performance status before establishing treatment goals. Certain outcome instruments have been validated for vertebral metastatic disease (EORTC, EQ-5D, QLOJL30, EQSD, HUI-MARK 3, ESAS, McGill Pain Questionnaire, VAS, MPAC). Appropriate surgical intervention is based on reasonable guidelines. The Harrington classification system combined with the Kostuik instability model, the Tomita grading system for vertebral tumor spread, and the Tokuhashi preoperative scoring system questionnaire are excellent guides for predicting outcomes of metastatic tumors and establishing methods for selecting the most appropriate treatment. The degree of the clinical validity of these guidelines has not been vigorously tested and, therefore, is lacking strong scientific support. Although they are not consensus- or evidence-based guidelines, they are useful tools for selecting a more appropriate surgical intervention. Furthermore, they can form the basis for establishing the standards for more widely acceptable guidelines.